Of course, there are many caveats. The current version is intended to support a particular experiment that will be run at George Mason University. The underlying code has more generality than is needed to support the market that the GMU economists are studying. You can probably learn a lot about our approach by reading the code, but you will also notice that I used a methodology based on Extreme Programming–I didn’t go back and clean up code as often as I would have earlier in my career. If you find things that could be improved in the code, by all means, send my your fixes.
Here are the release notes I uploaded with the tar files:
This is the first public release of the Zocalo Prediction Market code. There is both a source and a binary release here. The INSTALL file gives instructions for installing and running a simple experiment. RELEASE-TODO lists a few of the many tasks left to be done. This is very early in the life of Zocalo; there is much work left to do. The source release includes javadoc as well as the source code, and an ant file (build.xml) that will allow you to build source and binary releases, run junit tests, or regenerate the javadoc. The binary release includes just what you need to run a simple experiment, including the installation instructions.
LICENSE gives the CommerceNet license for the base Zocalo code. Third party code that we have included is described in THIRD_PARTY_SOFTWARE. The only third party code that we include source for is mod_pubsub.